Overview of China's Facial Recognition Industry Chain

Facial recognition, also known as face recognition, is a type of biometric technology that identifies individuals based on facial feature information. It primarily uses cameras or video streams to capture images containing faces, automatically detecting and tracking faces in the images, and then applying a series of related technologies to the detected faces. In recent years, with the development of artificial intelligence and the needs of national economic development and security defense, China’s facial recognition market has been continuously expanding, with technological levels steadily improving, achieving a world-leading position in algorithms. With the construction of smart cities and the development of smart homes, 3D facial recognition will have broader development prospects.

Industry Overview: Widely Used with Significant Advantages

Facial recognition can be broadly and narrowly defined. Broadly speaking, facial recognition includes a series of related technologies for building facial recognition systems, including facial image acquisition, facial localization, facial recognition preprocessing, identity confirmation, and identity retrieval; while narrowly defined, facial recognition specifically refers to the technology or system that verifies or retrieves identity through faces.

From the perspective of the industry chain, the upstream of the facial recognition industry is supported by hardware foundations, including high-definition cameras, processing chips (CPU, GPU, TPU), servers, and data and video transmission equipment. In terms of cameras, companies like Hikvision and Dahua are relatively advanced; in processing chips, international giants like NVIDIA, Intel, and AMD occupy a leading position; and in servers and data transmission, Chinese manufacturers like Huawei and Alibaba are developing rapidly.

The midstream of the industry chain mainly consists of facial recognition algorithms and software services. In terms of algorithm software, companies like Tencent, Baidu, Megvii Technology, CloudWalk Technology, and SenseTime are already at a global leading position; in hardware and software integration, security giants like Hikvision and Dahua, as well as Hanwang Technology and Sichuan University Intelligent Technology, are in fierce competition.

Currently, facial recognition in China has been widely applied in scenarios such as smartphone usage, identity verification at airports and train stations, remote account opening in finance, and corporate access control attendance. With the further maturity of technological conditions and the comprehensive arrival of the intelligent era, the application of facial recognition will further expand and deepen.

Facial recognition is part of biometric identification, which relies on physical characteristics of the human body for identity verification. These physical characteristics include fingerprints, voice, face, skeleton, retina, iris, and DNA, as well as behavioral characteristics such as signature actions, walking gait, and typing pressure. Among different biometric recognition methods, facial recognition has its unique characteristics, thus holding an important position in biometrics. Specifically, facial recognition has advantages such as non-invasiveness, no need for passive cooperation from the person, intuitiveness, and simplicity, which provide a foundation for the rapid development and application of facial recognition.

Market Status: Policy Boosting, Gradually Expanding Market Size

—— Development Accelerated After 2015.Facial recognition technology in China began in the late 1990s, going through five stages: technology introduction, professional market introduction, technology improvement, technology application, and usage across various industries. Among them, 2014 was a key year for deep learning applications in facial recognition, as FaceBook published an article titled “Deep Face: Closing in on Human-Level Performance in Face Verification.” After that, the teams of Yin Qi, founder of Face++, and Tang Xiaoou from the Chinese University of Hong Kong achieved remarkable results in the field of deep learning combined with facial recognition. Currently, domestic facial recognition technology has relatively matured and is being promoted across various fields.

—— Frequent Policies Boost Industry Development.With the development of science and technology, artificial intelligence (AI) has gradually gained attention from countries around the world, and research and promotion in the field of AI have been elevated to national strategic planning by many countries. As a subclass of AI, facial recognition has gradually found applications in various fields, with its precise identification characteristics prompting increased attention and promotion in various sectors. In recent years, the frequent introduction of relevant policies in China has also provided policy guarantees and support for the development of facial recognition technology.

In today’s information-exploding society, the security and concealment of information are becoming increasingly important. How to effectively and conveniently carry out identity verification and recognition is becoming a very prominent issue. Traditionally, various verification methods such as ID cards, password cards, and token cards have many problems, and with the continuous advancement of various cracking technologies, they face increasingly serious challenges. Facial recognition technology, as one of the most outstanding capabilities of human vision, due to its harmlessness and the most intuitive and natural way for users, has become the most promising method in the field of biometric automatic recognition technology. In recent years, China’s facial recognition industry market size has gradually expanded, exceeding 2.1 billion yuan in 2017, with a year-on-year growth of about 27%. In 2018, facial recognition technology was used in identity verification at several train stations in China, significantly increasing the market size to about 2.76 billion yuan, a 26% increase from 2017, and is in a rapid development stage.

—— Access Control Attendance Field Holds the Highest Proportion.With the development of technology and the increasing security requirements, facial recognition technology has undergone tremendous changes in industrial applications, rising from industries with lower security reliability requirements, such as security and communities, to industries with higher security reliability requirements, such as finance, social security, securities, banking, and internet finance. Currently, the application of facial recognition in attendance/access control is the most mature, accounting for about 42% of the industry market; security, as one of the earliest application fields for facial recognition, holds about 30% of the market share; finance, as one of the important future application fields for facial recognition, is gradually expanding its market size, currently accounting for about 20% of the industry.

Technical Level: Rapid Growth of Patent Applications, Chinese Algorithms Leading Globally

The scale of China’s facial recognition industry is rapidly expanding, backed by the rapid development of Chinese facial recognition technology. The wave of artificial intelligence has provided opportunities for the rapid development of facial recognition technology; the support of Chinese artificial intelligence policies has provided solid guarantees for the development of facial recognition. According to the latest data from Soopat, in recent years, the number of patent applications and disclosures for facial recognition technology in China has been continuously increasing. In 2017, the number of patent applications and disclosures for facial recognition in China was 3,368 and 2,698, respectively, and by 2018, the number of patent applications increased to 3,487, while the disclosure volume nearly doubled to 5,200.

The core of facial recognition technology is the algorithm, which is responsible for extracting facial features and comparing them with known faces in the database to complete the final classification, which is the most important process in the entire facial recognition. Currently, Chinese companies have achieved a globally leading position in the field of algorithms. According to the recognition rates of Chinese companies in the LFW (Labeled Faces in the Wild) database, the facial recognition technology of Ping An Technology under Ping An Group leads domestic and foreign well-known companies with a recognition accuracy of 99.84% and the lowest fluctuation, ranking first in the world; Tencent’s Youtu ranks second with a recognition rate of 99.80%; and Dahua ranks third with a recognition rate of 99.78%.

Note: The recognition rate is the result of facial recognition on outdoor photos in the LFW database under unconstrained and annotated conditions.

According to the latest results from the National Institute of Standards and Technology (NIST) global facial recognition algorithm test (FRVT) in 2018, a total of 39 companies and institutions from around the world participated in this competition. In the latest ranking, the top five algorithms were all from Chinese companies, demonstrating the strong competitiveness of Chinese companies. Among them, Yitu Technology’s algorithms took the top two places, SenseTime took third and fourth places, the Shenzhen Institute of Advanced Technology of the Chinese Academy of Sciences took fifth place, and Megvii Technology’s algorithm took eighth place.

Note: The error rate is the result of facial recognition under FMR<0.000001 conditions in the FRVT visa photo recognition results.

Both the LFW test results and the NIST test results demonstrate that China’s facial recognition technology is currently at a global leading position. With the improvement of China’s overall technical level and the expansion of investment in artificial intelligence fields such as facial recognition, China is likely to continue to lead the world in the future.

Investment Situation: Capital Pursuit, Startups Frequently Receive Large Investments

Facial recognition technology can be widely applied in the commercial field and can generate significant market returns, making companies that master core technologies in facial recognition the target of venture capital. The demand for information technology development and intelligent development in China has given rise to a batch of promising artificial intelligence startups. Benefiting from government policies, many startups focus on facial recognition and have received significant investments in recent years through close cooperation with the government.

In August 2012, Face++ (Megvii Technology) received angel financing from Lenovo Star before its establishment, and the following year it received millions of dollars in Series A financing from Innovation Works, followed by $22 million in Series B financing in November 2014, with investors including Qiming Venture Partners and Innovation Works. At the same time, startups in facial recognition such as SenseTime, Megvii Technology, and Yitu Technology have also received large investments, with SenseTime receiving $600 million and $620 million in April and May 2018, respectively, and $1 billion from SoftBank China in October.

Competitive Landscape: Different Fields, Each with Its Own Strengths

Currently, the leading facial recognition companies in China mainly include Megvii Technology and SenseTime, which are startups specifically established for research and application of facial recognition technology, as well as established security and listed companies such as Hikvision, Dahua, Sichuan University Intelligent Technology, and Obit, along with internet giants Tencent, Alibaba, and Baidu entering the facial recognition field either through investment or their own R&D teams.

In different application fields, the brand awareness of facial recognition industry varies. According to the application dimensions of facial recognition technology, it can be divided into government, enterprise, and individual consumers. Government departments generally hope to apply facial recognition technology in smart security fields, where application scenarios are complex and require high accuracy; individual consumers have lower complexity in application scenarios but have higher requirements for consumer experience. According to the supply dimension of facial recognition technology, the products that facial recognition technology can provide are mainly divided into engineering projects, hardware, and software technologies.

In the personal application field, pure software technology (facial recognition technology) is combined with smart devices such as smartphones and tablets, where application scenarios are simple, mainly dominated by startups like Megvii Technology and SenseTime; in the enterprise application field, the main product demand is for access control and attendance, where application scenarios are the simplest, mainly dominated by companies like Hanwang Technology and Hikvision; in the government application field, facial recognition project engineering is generally applied in public safety fields (including entry and exit management, smart city, etc.), where application scenarios are the most complex, mainly involving companies like Obit, Haixin Technology, Hikvision, and Dahua.

Development Prospects: Driven by Multiple Factors, Market Size Steadily Expanding

As facial recognition technology continues to mature, market demand will accelerate, and application scenarios will be continuously explored. From social security benefits to campus access control, from remote pre-authorization to security gate inspections, facial recognition is continuously opening up markets. The high heat of the facial recognition market and the leapfrog development of its application scenarios are fundamentally due to technological innovation. Under artificial intelligence, deep learning has allowed the accuracy of facial recognition to surpass human-level performance, greatly enriching the application scenarios of facial recognition. The essential demand for remote account opening in internet banking has brought facial recognition into financial-grade application scenarios, while the construction of smart cities has gradually expanded the demand for facial recognition in security and other fields; frequent layouts by giants in facial recognition grant it greater imaginative space for application scenarios, while cultivating user habits of “face scanning” and recognition of technology, which is conducive to further industry development. The multilateral push has allowed facial recognition applications to experience explosive growth.

With the steady development of China’s social economy, the demand for the application of facial recognition technology will also increase. In addition, with the in-depth development of domestic safe city and smart city projects, the high-definition nature of urban monitoring has further become popularized, with the number of cameras increasing significantly, greatly reducing the obstacles to data collection for facial recognition, enhancing the quality and application fields of facial recognition. It is expected that in the next five years, the facial recognition market will maintain a growth rate of over 20%, reaching a market size of about 10 billion yuan by 2024.

The above data comes from: Qianzhan Industry Research Institute

Outstanding Companies and Products in Facial Recognition:

Beijing SenseTime Technology

Core Business: Smart Finance, Smart Business, Smart Security, Internet+

Technical Features: Facial Technology, Intelligent Monitoring, Image Recognition, Text Recognition, Image and Video Editing, Deep Learning Framework

Successful Cases: JD.com, China Merchants Bank, Jiedaibao, Lakala, Great Wall Club, Oriental Net Power, Public Security Third Institute, NVIDIA, iFLYTEK, etc.

Notes: Collaborations with JD.com, UnionPay, China Merchants Bank, Lakala, Rong360, etc.; layout of smart city security projects; the SenseFace facial control system has begun to be widely implemented; the image search system has been applied in public security bureaus in Guangzhou, Chongqing, Hebei, etc.; the SenseAR augmented reality engine application; the background blur function of portraits, and the facial clustering function in smart albums are applied in OPPO, Xiaomi, and other smartphones.

Official Website: https://www.sensetime.com/

Beijing Megvii Technology

Core Business: FaceID online identity verification service, Face++ AI open platform, smart real estate solutions, smart security solutions

Technical Features: Dynamic facial recognition, online/offline liveness detection, real-time retrieval from large portrait databases, document recognition, pedestrian detection, trajectory analysis

Successful Cases: CITIC Bank, China Merchants Bank Credit Card Center, Ant Financial, Shenzhou Car Rental, Cloud End Wuhan, Neusoft Group, CapitaLand Group, 2017 Boao Forum for Asia, etc.

Notes: Provides facial login functionality support for Alipay client; provides facial recognition technology support for the Ministry of Public Security’s First Research Institute; provides technical support for Meitu’s products; applies facial recognition technology to verify driver identities in eDaijia, Yidao, and Shenzhou Car Rental; Megvii’s intelligent open platform Meagvii Cloud is an AI open platform that provides developers with facial recognition, text recognition, image recognition, and other AI capabilities.

Official Website: https://megvii.com/

Guangzhou CloudWalk Technology

Core Business: Rapid deployment platform, intelligent hardware, intelligent systems

Technical Features: Facial information detection, feature extraction and fusion, quality attribute analysis, facial similarity comparison, feature similarity comparison, OCR recognition

Successful Cases: Agricultural Bank of China, LeEco Car Networking, Guangdong Provincial Public Security Department, Guangdian Yuntong, Northwest Star, Yuyin Co., Beijing West Railway Station, Heilongjiang Reclamation Area, Shandong Urban Commercial Bank Alliance, etc.

Notes: Invited to draft and formulate national standards for facial recognition; super counters at Agricultural Bank of China, facial recognition for withdrawal; security products have been launched in 22 provinces for practical use.

Official Website: http://www.cloudwalk.cn/

Shanghai Yitu Technology

Core Business: Intelligent security platform, urban data brain, smart medical health, intelligent hardware devices

Technical Features: Computer vision, natural language understanding, knowledge reasoning, intelligent hardware, robotics technology

Successful Cases: Chinese Customs, China Border Inspection, Huawei, Alibaba Cloud, China Merchants Bank, Wanda Group, Suzhou Public Security Department, Fujian Public Security Department, Guizhou Public Security Department, etc.

Notes: For China Merchants Bank, Pudong Development Bank, JD Finance, 360 Jin Control; Jiangsu Provincial Public Security Department uses Yitu system.

Official Website: http://www.yitutech.com/

Baidu

Baidu is using facial recognition turnstiles internally; in November 2016, it cooperated with the Wuzhen Scenic Area, allowing tourists to freely enter and exit the scenic area with facial recognition; signed an agreement with Capital Airport, aiming to achieve facial recognition boarding in the future; partnered with the “Baby Home” public welfare platform to use facial recognition to find missing children; collaborated with Rainor Co., through service integrator Cella, to jointly provide smart pharmacy solutions for the pharmaceutical retail industry, which have been applied in Xiansheng Zhi Kang chain pharmacies.

Official Website: https://www.baidu.com/

Alibaba

Facial recognition technology modules can be freely combined through API parameters, with flexible service customization; based on deep learning and massive facial annotation data, combined with Alibaba Cloud’s technical strength, it can provide stable and reliable high-traffic services; with facial recognition, high efficiency and accuracy can screen products that use celebrities’ endorsements without permission, thereby ensuring the advertising effectiveness of celebrity-endorsed products in Alibaba’s direct marketing and display ads.

Official Website: http://www.alibaba.com/

Tencent

Tenpay has reached a strategic cooperation for image comparison services with the National Citizen ID Number Query Center under the Ministry of Public Security; Youtu’s facial recognition technology will be widely applied in the express delivery of government affairs, valuable items, and important documents; applied in Tencent’s WeChat securities and other products.

Official Website: http://www.qq.com/

Beijing Hanwang Technology

Core Business: Electronic products related to individuals and families, electronic products related to enterprises and offices

Technical Features: Handwriting technology, OCR technology, keyboard technology, formula recognition, input method

Successful Cases: Agricultural Bank of China, China Life, Acer, TCL, BuBuGao, Dopod, HP, Haomijixing, People’s Publishing House, etc.

Notes: Assisting the Yinchuan Municipal Government in applying biometric technology to build a smart government platform; helping the Hangzhou National Tax Bureau realize facial biometric recognition technology for tax payment; widely applied in public security criminal investigation and fugitive tracking, technology licensing has been reached with ASUS, Haier, Changhong, Hisense, Ping An Bank, etc., and promoted to smart appliances, laptops, mobile terminals, and other application platforms.

Official Website: http://www.hw99.com/

Hefei iFlytek

Core Business: Voice engine, educational products, mobile applications, interactive music

Technical Features: Speech recognition and synthesis technology, natural language processing technology, speech evaluation technology, voiceprint recognition technology, handwriting recognition technology

Successful Cases: iFlytek Input Method, Lingxi Recording Treasure, Happy Bear, Listening and Speaking Without Worry, iFlytek TV Assistant, iFlytek Smart Speaker, etc.

Notes: iFlytek, in collaboration with Professor Tang Xiaoou’s team from the Chinese University of Hong Kong, jointly launched world-leading facial recognition technology, providing facial verification, online/offline facial detection, and facial key point detection functions; in collaboration with China UnionPay and WeBank, launched the “voiceprint + face” fusion authentication personal transfer application, iFlytek is using identity authentication attendance, and branches across the country can use the APP to check in with “face + voiceprint”.

Official Website: http://www.iflytek.com/

Sichuan Sichuan University Intelligent Technology

2D facial recognition products have been launched in the market, while 3D facial collection and recognition products are mainly in the engineering prototype and product prototype stage, with the main application field being public safety. The 2D product prototype system has been applied in student dormitories at Beijing Normal University and Sichuan University, and is being tested in railway ticket verification at Chengdu Railway Station.

Official Website: http://www.wisesoft.com.cn/

Shanghai Yuedian Technology

“Yueke” is a software-integrated customer analysis terminal, and “Yuelin” provides an integrated intelligent access control and facial recognition solution.

Official Website: https://www.readsense.cn/

Beijing Orion Star

In addition to access control, it is also applied to everyday scenarios such as mobile phones, and is used in Cheetah Mobile’s live streaming product Live.me, which includes backend technology detection and identification of pornographic information, recognition of official website logos for advertising detection, and classification of hosts based on tags such as gender and type, achieving dynamic facial mapping and other personalized functions.

Official Website: http://www.ainirobot.com/

Beijing Geling Deep Vision

Launch of the deep vision camera FOVEACAM aimed at public security and transportation industries, capable of recognizing faces at long distances.

Official Website: http://www.deepglint.com/

Beijing ZK Aosen

DeepEyes dual-camera deep learning facial recognition anti-counterfeiting technology.

Official Website: http://www.authenmetric.com/

Shenzhen Ping An Technology

Facial recognition technology has been applied in more than 10 scenarios, such as remote account opening, card binding verification, account login, installment shopping, facial attendance, and facial payment in over 50 terminal applications.

Beijing Haixin Zhisheng

In 2010, completed the facial recognition monitoring and alarm system as the core for the Shanghai World Expo park’s facial collection and comparison system construction project; in 2012 became the contractor for the national citizen ID number query service center’s portrait comparison authentication system, supporting a capacity of over 1.1 billion people; in 2016, introduced three identity verification devices for the public recruitment examination of Shijingshan District, allowing candidates to enter the examination venue with facial recognition.

Official Website: http://www.biometric.net.cn/

Beijing Feisou Technology

Has developed multiple online and offline SDK core product lines for enterprises, including facial detection, facial landmark positioning, facial recognition, celebrity recognition, facial attribute recognition, and target/scenario recognition, as well as pornographic image recognition.

Official Website: http://www.faceall.cn/

Tianjin Hanbo Technology

Facial recognition products have been launched, including access control (GATES), recognition terminals (DOORS), turnstiles (INS), desktop terminals (ONS), and a series of industry solutions.

Official Website: http://www.opzoon.com/

Zhuhai Renren Intelligent

Released facial recognition hardware modules, aceos, developed based on ARM chips, can be easily embedded into smart cameras and other terminals, mainly applied to identity verification and facial comparison in security systems.

Official Website: http://www.openface.com.cn/

Suzhou Zhongde Hongtai

“Deep Eye” provides a complete set of identity recognition technology solutions, including facial recognition.

Official Website: http://www.vdin.com.cn/

Hunan Vision Weiye

Core Business: Intelligent security platform, urban data brain, smart medical health, intelligent hardware devices

Technical Features: Intelligent recognition system for people and vehicles, facial recognition, image enhancement processing engine, automatic aging estimation of faces, computing architecture from frontend to cloud

Successful Cases: Tiananmen Square, Capital Airport T3 Terminal, Beijing Subway, Beijing Public Security, Changsha Huanghua Airport, Hunan Provincial Party Committee, Changsha Pozi Street, Nanjing University, Zhejiang University, etc.

Hangzhou Hikvision

Core Business: Monitoring equipment, camera equipment, vehicle-mounted products, software services, etc.

Technical Features: SMART IR technology, ARM and DSP architecture, motion detection technology, embedded system priority inversion, audio and video codec technology

Successful Cases: Bird’s Nest, Olympic Village, World Expo, Qinghai-Tibet Railway, Shaanxi Industrial and Commercial Bureau Administrative Building, Ocean Group, Nanjing Gulou Hospital, etc.

Official Website: http://www.hikvision.com/

Jiadou Technology

Core Business: Smart city, intelligent rail transit, service and product integration

Technical Features: Dynamic facial recognition technology, automatic ticketing system, barrier gate system, comprehensive monitoring system, video monitoring system

Successful Cases: Guangzhou High-speed Highway, Shunde District Moat, Baise Electronic Police, Zhongshan Public Security Police, Guangdong Provincial Guard Comprehensive Command and Dispatch, Guangzhou Online Vehicle Management Office, Ping An Asian Games, Qingdao South District “Sky Net” project, etc.

Guangdong Boya Information

Core Business: Facial recognition, intelligent transportation systems, video and environmental monitoring equipment, access control systems, recognition deployment products, photo collection instruments, intelligent analysis, platform software

Technical Features: Liveness detection, identity verification, real person verification, 1:1 facial comparison, 1:N facial comparison, N:N facial comparison, real-name verification

Successful Cases: Wing Payment, Huawei, Microsoft, New Wave, Yushi Technology, HP, Lenovo, Shenzhou Digital, Baode, etc.

Official Website: http://www.poya.com.cn/

Beijing Zhongke Shituo

With facial recognition and deep learning as the core, it provides various professional technical services in the field of computer vision and machine learning for government, enterprise, and commercial clients, offering integrated solutions for hardware and software, assisting clients across industries to gain autonomous iteration and self-learning capabilities in the era of artificial intelligence.

Official Website: http://www.seetatech.com/

Shensi Electronics

A leading provider and service provider of identity recognition solutions in China, implementing a strategic upgrade from “deep cultivation in the industry to industry integration, from identity recognition to intelligent cognition,” continuing to provide one-stop services with artificial intelligence characteristics in identity recognition-related fields.

Official Website: http://www.sdses.com/

Gaoxin Technology

A global leader in smart city IoT products and services. Since its establishment in 1997, the company has focused on the research and development of products and technologies related to perception, connection, and platform layers based on IoT architecture, deeply cultivating the markets of smart cities and vertical industries, focusing on safe cities, vehicle networking, public security information technology, railway safety, financial security, and other fields.

Official Website: http://www.gosuncn.com/

ColorReco

A technology innovation company that builds a new generation of facial recognition technology through deep learning frameworks, with core technologies that have complete independent intellectual property rights, with original innovations in facial detection, recognition, attribute detection, liveness detection, and facial search. Based on the facial recognition needs in different business scenarios, it provides industry-leading visual identity authentication technology and overall solutions for enterprise-level users and partners. Committed to building a new generation of computer vision understanding and artificial intelligence engines, enabling computers and other intelligent hardware to understand our world.

Official Website: http://www.colorreco.com/

Linkface

A technology innovation company focused on computer vision research. The company aims to “reshape the singularity of the financial world” as its vision, using deep learning as its technical engine, combining computer vision and big data analysis, dedicated to providing automated solutions for identity verification, text information recognition, basic data query, user profiling, and anti-fraud for traditional banks, internet finance, insurance, and securities institutions through original image recognition algorithms and big data processing technologies, helping financial clients build safer, more efficient, and smarter service experiences.

Official Website: https://www.linkface.cn/

Hunan Yuanjun Technology

Mainly engaged in research and product application development of facial recognition, behavior recognition, and robotic visual analysis and detection core algorithms, with military and social public safety fields as the main application direction, selling self-developed hardware and software products, software customization, supporting information system integration and operation management services, etc., providing integrated overall solutions for artificial intelligence. Main products: embedded facial recognition channel control systems, three-in-one multi-biometric real-name verification products, non-perceptual facial recognition systems, abnormal behavior recognition systems, as well as facial recognition SDK sales and secondary development services based on specific user needs.

Official Website: http://www.hnfar.com/

Source:ittbank

Overview of China’s Facial Recognition Industry Chain

Leave a Comment Cancel reply